US regulators are delaying cyber resilience tests for banks to give them time to harden defenses against AI-powered threats, including deepfake fraud and voice cloning attacks that are reshaping financial cybersecurity.
An AI-powered voice system at Glendale Community College's commencement mispronounced and skipped graduate names, exposing limitations in synthetic voice deployment for high-stakes live events.
YouTube is rolling out its likeness detection feature globally, letting creators identify and request removal of AI-generated deepfake videos that use their face or voice without consent.
Reality Defender publishes an enterprise incident response playbook for deepfake attacks, offering CISOs structured guidance on detection, containment, and recovery as synthetic media fraud escalates across financial services and corporate environments.
ZAYA1-8B, a Mixture-of-Experts model with just 760M active parameters trained entirely on AMD GPUs, reportedly outperforms GPT-5-High on math benchmarks—challenging NVIDIA's dominance in frontier AI training.
A North Korea-linked threat group is deploying AI-generated deepfakes in spear-phishing operations against South Korean officials, marking another escalation in nation-state weaponization of synthetic media for espionage.
Elon Musk's xAI has launched Grok Build, an AI coding agent aimed at competing with Anthropic's Claude Code and OpenAI's Codex. The move escalates the battle for developer mindshare in agentic coding tools.
Bug bounty platforms are being inundated with AI-generated vulnerability reports that look credible but describe non-existent flaws, forcing triage teams to waste hours debunking hallucinated security claims.
Financial institutions face a surge in deepfake-driven fraud as synthetic voice and video attacks bypass traditional KYC and identity verification systems, forcing banks to overhaul authentication strategies.
Training generative models typically demands massive datasets, but transfer learning offers a path forward with limited data. Here are six proven techniques researchers use to fine-tune GANs and diffusion models efficiently when training samples are scarce.
ArXiv is tightening rules on AI-generated submissions, threatening year-long bans for authors who outsource entire papers to large language models — a major shift in how the research repository handles synthetic content.
GetReal Security is doubling down on deepfake detection and human verification, addressing identity risks emerging from generative AI's rapid expansion into voice, video, and image synthesis.